Integrating principal component analysis and optimal histogram estimation for Bayesian control loop diagnosis

Control loop performance monitoring and diagnosis is an important practical topic in process industries. Data-driven Bayesian method is attracting more and more research attention, which uses multiple monitors to yield probabilistic assessments. However, the correlation among variables, high-dimensional observations, and discrete degree of the evidence affects the diagnostic performance. This paper presents a novel Bayesian method which integrates principal component analysis (PCA) and optimal histogram estimation (OHE) for efficient multimode fault diagnosis. PCA is firstly applied to transform the original observations into uncorrelated principal components (PCs). Then OHE is implemented to select optimal bin width used in discrete evidence in order to improve the accuracy of likelihood estimation. Finally, the performance of OHE-PCA Bayesian diagnosis is examined through Tennessee Eastman (TE) benchmark process. The comparison demonstrates the feasibility and effectiveness of the proposed method.

[1]  Biao Huang,et al.  Bayesian Control Loop Diagnosis by Combining Historical Data and Process Knowledge of Fault Signatures , 2015, IEEE Transactions on Industrial Electronics.

[2]  D. W. Scott,et al.  Multivariate Density Estimation, Theory, Practice and Visualization , 1992 .

[3]  Biao Huang,et al.  Control Performance Assessment Subject to Multi-Objective User-Specified Performance Characteristics , 2011, IEEE Transactions on Control Systems Technology.

[4]  Hailin Li,et al.  Accurate and efficient classification based on common principal components analysis for multivariate time series , 2016, Neurocomputing.

[5]  Sirish L. Shah,et al.  Performance Assessment of Control Loops: Theory and Applications , 1999 .

[6]  Biao Huang,et al.  Multiple oscillations detection in control loops by using the DFT and Raleigh distribution , 2015 .

[7]  Biao Huang,et al.  Expectation–Maximization Approach to Fault Diagnosis With Missing Data , 2015, IEEE Transactions on Industrial Electronics.

[8]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part II: Qualitative models and search strategies , 2003, Comput. Chem. Eng..

[9]  Fei Qi,et al.  Bayesian approach for control loop diagnosis , 2011 .

[10]  Chen Jing,et al.  SVM and PCA based fault classification approaches for complicated industrial process , 2015, Neurocomputing.

[11]  Haitao Zhao,et al.  Multimode Process Monitoring Based on Sparse Principal Component Selection and Bayesian Inference-Based Probability , 2015 .

[12]  Anna Pernestål,et al.  A Bayesian Approach to Fault Isolation with Application to Diesel Engine Diagnosis , 2007 .

[13]  E. F. Vogel,et al.  A plant-wide industrial process control problem , 1993 .

[14]  Ping Zhang,et al.  A comparison study of basic data-driven fault diagnosis and process monitoring methods on the benchmark Tennessee Eastman process , 2012 .

[15]  J. Edward Jackson,et al.  Quality Control Methods for Several Related Variables , 1959 .

[16]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part III: Process history based methods , 2003, Comput. Chem. Eng..

[17]  Raghunathan Rengaswamy,et al.  A review of process fault detection and diagnosis: Part I: Quantitative model-based methods , 2003, Comput. Chem. Eng..

[18]  Biao Huang,et al.  Control loop diagnosis with ambiguous historical operating modes: Part 2, information synthesis based on proportional parametrization , 2013 .

[19]  Theodora Kourti,et al.  Statistical Process Control of Multivariate Processes , 1994 .

[20]  C. T. Seppala,et al.  A review of performance monitoring and assessment techniques for univariate and multivariate control systems , 1999 .

[21]  S. Joe Qin,et al.  Control performance monitoring — a review and assessment , 1998 .

[22]  T. Harris.  Assessment of Control Loop Performance , 1989 .

[23]  Biao Huang,et al.  A Bayesian approach for control loop diagnosis with missing data , 2009 .